Yes, please leave email the way it is, because it works fine. It’s one of the few things that does.
The only feature I want is end to end encryption. Google will never make encryption easy on their own because they’ve got perverse incentives not to.
Email is one of the last widely used open communication protocols, and it is sad that people are so resistant to its evolution because that is exactly what is accelerating its demise in favor of proprietary software.
I agree that SSL in transit and GPG could stand to be improved, but they do exist.
> No server identity verification.
What do you want that SPF+DKIM doesn't cover?
> No delivery confirmation.
Good? That sounds like asking for abuse. And in any event, it's no different than inline remote images (which of course are blocked for the same reason that delivery confirmation is a bad idea). Unless you just mean being sure that the message got to the target server, in which case I agree that the best we can do is delivering it directly to the target MX and see that it confirmed that it received the message.
> No push standard.
Temporarily true, pending JMAP (which I suppose strengthens your point about needing innovation).
> No concept of groups or other ways to categories and manage access between users in a domain.
Not quite sure what this means. Do you want something like shared mailboxes and tags?
> Every provider has their own implementation of threaded conversations.
True, though in a message-centric format I'm not sure that it's a bug.
> No major provider supports non-ASCII addresses.
Not a bug in email, only arbitrary restrictions by certain providers. Also opens you up to unicode normalization attacks (phishing is easier if you can fake characters).
> Servers can't even agree on simple message formatting.
I'm not sure what formatting servers even care about (email is just some blobs of text strung together), but I can't really refute this without knowing what you refer to more specifically.
> Most things that make email actually useful are outside the spec and tacked on by each individual provider.
Email is useful because it lets us send arbitrary stuff (usually a text body and zero or more attachments) between federated providers. It's a simple-ish protocol that is stable and just works.
EDIT: In summary, I vastly prefer email as it is: A simple protocol that can include arbitrary information, which allows people to extend it however they want. Of course, this means that extensions are arbitrary and at best non-required, but that keeps it flexible and back-compatible.
Simple tools that do one thing well are almost always best.
What we need is a way to make E2E encryption easy and straightforward, like Signal or Wire, where it's easy, automatic and default enabled.
Hard to separate pending tasks from just conversations.
Hard to split conversations into separate threads when two topics diverge.
Leaving a conversation isn't just your decision: you have to both reply all and hope that no one else replies all using a previous email in the thread.
In a corporate email, inability to prevent mass emails from other employees. Hard to do for personal email too.
EDIT: I'm a huge fan, e.g., of the "travel bundles" Inbox creates. It's a giant win in usability for the use case of getting confirmation emails of booked trips. But you can see how it's a battle uphill that Inbox fights against the protocol, as the AI isn't always able to bundle all relevant information together, or extract key pieces of info from the email bodies.
